The disclosures herein relate generally to a computer system and more particularly to a speaker in a computer chassis.
When a speaker is mounted inside a computer chassis without a speaker box, the resulting sound performance is poor due to the speaker sound energy being radiated inward and outward of the chassis. When a speaker box with a flat rear wall is employed to solve this problem, the reflected sound off the flat rear wall will have one base resonance frequency, causing a peak in the perceived speaker volume as the driving frequency increases.
Currently, a common approach to this problem is to use a soft material, such as foam, on the rear wall of the speaker box in order to reflect the sound back at various frequencies. This solution does vary the natural responses and avoids a peak in the perceived speaker volume, but it also absorbs overall sound energy which reduces sound quality and is relatively costly.
Therefore, what is needed is a speaker box that varies the natural responses and avoids a peak in the perceived speaker volume, but does not absorb sound energy and is less costly to manufacture.
One embodiment, accordingly, provides multiple resonant frequency lengths between the speaker and a hard reflecting surface. To this end, a speaker apparatus includes a speaker and an acoustic box connected to the speaker. The box has a sound reflecting distal wall including a plurality of stepped portions of variable distances from the speaker.
A principle advantage of this embodiment is that it retains the advantages of a hard reflecting surface while being less costly than attaching a soft material to the rear speaker box wall. The stepped multi-leveled rear wall spreads the natural responses of the speaker and avoids a peak in the perceived speaker volume without absorbing as much sound energy as a soft material attached to the back of the speaker box.